"Tabloids are easier to read than broadsheets"
Mathematics Coursework - Data Handling
"Tabloids are easier to read than broadsheets"
Specify the Problem and Plan
I am going to investigate whether a tabloid paper is easier to read than a broadsheet.
* Tabloid - 'A newspaper of small format giving the news in condensed form, usually with illustrated, often sensational material.'
* Broadsheet - 'A size and format for newspapers, and a descriptive term applied to papers which use that format rather than the smaller tabloid format.'
Definitions taken from: http://www.answers.com/
I am going to collect will be primary. As the data, I am going to collect it reasonably accurate and secondary data is not needed. I might have done a public vote using questionnaires but that would be a biased data and that is accessing the public favouritism but I am going to investigate readability not the favourite paper among the public. Therefore, I am going to use a mathematical way of calculating the readability of one of each type of paper and compare them, and come to a conclusion which paper is easier to read.
Readability can be measured by using several factors including:
* Word Length
* Sentence Length
* Vocabulary sophistication
* Sizes of Paragraph
* Percentage of Words to Pictures ratio on the page
* Length of the Article
* Size of Writing
* Microsoft Word - readability analysis
The data I am going to use to get my data from to compare the two papers are, one broadsheet and the other tabloid, which will be both dated on the same day to get familiar article from both papers allowing it to be a fair test. I will be using:
* Broadsheet - The Times newspaper - Thursday 17th February 2005
* Tabloid - The Sun newspaper -Thursday17th February 2005
Fair Test Precautions
. I will pick the articles from both papers from the same section e.g. sports, politics or gossip. So that the articles will be roughly similar.
2. I will two papers from the same date to ensure that the articles are similar.
3. I will also conduct a pre-test and from this I could then rectify so that my main test could be fair and enable me to make a prediction based on evidence
4. From my pre-test, which I will conduct, I will learn which certain non- word items if I should count them as words when counting the amount of words.
5. I will use a sampling method to make sure I get a range throughout the whole article
6. I will use more than one test for readability to be more accurate, several factors can be used to determine readability of the newspaper, as all the factors are equally valid when measuring readability It will also eliminate a few of the bias factors and the analysis of my data will more likely fit with my hypothesis.
7. I will use more than one article from the both papers to compare to each other from different sections.
Corresponding to all these fair test precautions into my investigation, I have eliminated all bias, limitations to my work; I should be able to reach a balanced conclusion.
Word Length
The first test for readability I am going to use is word length. I will count the frequency of the words of an article and count the number of letters in each word. This factor is related to readability, as more complicated and longer words are harder to read and than easier words.
Pre Test
For the pre test I used I compared the one article from each paper. In general, I realised that the broadsheet newspaper articles were longer in length than tabloid newspapers articles. For this reason, I decided to have a stratified sample so that the sample would be in proportion to the length of the article. I did a pre test to help me define my variables and my values I realized that my strata of 40 was accurate and will be a very accurate base to base my conclusion on. From my pre-test I now ...
This is a preview of the whole essay
Pre Test
For the pre test I used I compared the one article from each paper. In general, I realised that the broadsheet newspaper articles were longer in length than tabloid newspapers articles. For this reason, I decided to have a stratified sample so that the sample would be in proportion to the length of the article. I did a pre test to help me define my variables and my values I realized that my strata of 40 was accurate and will be a very accurate base to base my conclusion on. From my pre-test I now know what to count as a word a list below shows what I will and won't except as a word which I will count.
Will
* Abbreviations
* Words with forward slash = 2 words
* Hyphened words = 1 word
* Websites / Email addresses
* Names
Won't
* Numbers (figure format)
* Punctuation
* Words with apostrophe as two words
Stratified Sampling
I will count the number of words in each article and to get the total number and then divide it by 40 whatever my result I will count that many words and highlight the word this will allow me then to go back and calculate the word length. Using a Stratified sampling method I will get a sample from the whole text.
Hypothesis
I predict that the average word length in a broadsheet will be greater than a tabloid, as a broadsheet uses far more complex and informal vocabulary but the Inter-quartile range for the tabloid will be greater but further down the graph. Therefore, as readability is considered I think a tabloid will be easier to read than a broadsheet, as the word length would be longer in a broadsheet, which uses more complex and informal language, which you would expect to have a longer word length.
Percentage of Words to Pictures ratio
On the pages of the paper, I will measure the percentage of text on a page and the percentage of pictures by measuring the area covered (in cm2). It is assumed that with bigger pictures in relation to the text are much easier to read, as the pictures are easier to understand and interpret.
Pre test
For my Pre test I compared one article from each paper and, I have decided not to look at the whole page but instead look at specific articles and measure how much of it is pictures and the actual text, but I will still measure it the same way. From my Pre-test I have decided to count captions and titles and sub headings as pictures and text as text, this will allow me to see how much text is actually on the article compared to the rest of it.
Hypothesis
I predict that there will be greater percentage of writing compared to the picture in a broadsheet than a tabloid, because broadsheet papers write in more depth and therefore need more space for text rather than photos. I also think the ratios will differ depending on the article topic and where it is in the paper.
Collect, Process and Represent
Word Length
Front Page
Tabloid - The Sun
Broadsheet - The Times
Length (x)
Freq.
Percentage (f)
C.F (%)
fx
Length (x)
Freq.
Percentage (f)
C.F (%)
fx
3
8.57%
8.57%
8.5714
2
5.13%
5.13%
5.1282
2
6
7.14%
25.71%
34.286
2
7
7.95%
23.08%
35.897
3
9
25.71%
51.43%
77.143
3
9
23.08%
46.15%
69.231
4
2
5.71%
57.14%
22.857
4
5
2.82%
58.97%
51.282
5
5
4.29%
71.43%
71.429
5
4
0.26%
69.23%
51.282
6
3
8.57%
80.00%
51.429
6
3
7.69%
76.92%
46.154
7
3
8.57%
88.57%
60
7
3
7.69%
84.62%
53.846
8
2
5.71%
94.29%
45.714
8
0
0.00%
84.62%
0
9
2.86%
97.14%
25.714
9
2.56%
87.18%
23.077
0
0
0.00%
97.14%
0
0
2
5.13%
92.31%
51.282
1
2.86%
00.00%
31.429
1
3
7.69%
00.00%
84.615
2
0
0.00%
00.00%
0
2
0
0.00%
00.00%
0
?
35
00.00%
00.00%
428.57
?
39
00.00%
00.00%
471.79
Mean =
4.2857
Mean =
4.7179
Olympic Bid Article
Tabloid - The Sun
Broadsheet - The Times
Length (x)
Freq.
Percentage (f)
C.F (%)
fx
Length (x)
Freq.
Percentage (f)
C.F (%)
fx
2
5.00%
5.00%
5
0
0.00%
0.00%
0
2
9
22.50%
27.50%
45
2
4
0.26%
0.26%
20.513
3
4
0.00%
37.50%
30
3
9
23.08%
33.33%
69.231
4
6
5.00%
52.50%
60
4
7
7.95%
51.28%
71.795
5
7
7.50%
70.00%
87.5
5
7
7.95%
69.23%
89.744
6
6
5.00%
85.00%
90
6
5
2.82%
82.05%
76.923
7
2
5.00%
90.00%
35
7
3
7.69%
89.74%
53.846
8
2
5.00%
95.00%
40
8
2
5.13%
94.87%
41.026
9
2.50%
97.50%
22.5
9
2.56%
97.44%
23.077
0
0
0.00%
97.50%
0
0
0
0.00%
97.44%
0
1
0
0.00%
97.50%
0
1
0
0.00%
97.44%
0
2
2.50%
00.00%
30
2
2.56%
00.00%
30.769
?
40
00.00%
00.00%
445
?
39
00.00%
00.00%
476.92
Mean =
4.45
Mean =
4.7692
Liverpool Manager Article
Tabloid - The Sun
Broadsheet - The Times
Length (x)
Freq.
Percentage (f)
C.F (%)
fx
Length (x)
Freq.
Percentage (f)
C.F (%)
fx
4
0.00%
0.00%
0
0
0.00%
0.00%
0
2
1
27.50%
37.50%
55
2
0
25.64%
25.64%
51.282
3
4
0.00%
47.50%
30
3
9
23.08%
48.72%
69.231
4
9
22.50%
70.00%
90
4
8
20.51%
69.23%
82.051
5
3
7.50%
77.50%
37.5
5
4
0.26%
79.49%
51.282
6
4
0.00%
87.50%
60
6
2.56%
82.05%
5.385
7
2
5.00%
92.50%
35
7
3
7.69%
89.74%
53.846
8
2.50%
95.00%
20
8
2.56%
92.31%
20.513
9
2.50%
97.50%
22.5
9
2.56%
94.87%
23.077
0
2.50%
00.00%
25
0
2
5.13%
00.00%
51.282
1
0
0.00%
00.00%
0
1
0
0.00%
00.00%
0
2
0
0.00%
00.00%
0
2
0
0.00%
00.00%
0
?
40
00.00%
00.00%
385
?
39
00.00%
00.00%
417.95
Mean =
3.85
Mean =
4.1795
Percentage of Text to the Rest of the Article
Front Page
Olympic Bid
Liverpool Manager
The Times
The Sun
The Times
The Sun
The Times
The Sun
Text
71%
6%
55%
8%
94%
5%
Rest of the Article
29%
94%
45%
92%
6%
85%
Interpreting and Discussing the Results
Word Length
In my hypothesis, I predicted that the average word length in a broadsheet would be greater than a tabloid, as a broadsheet uses far more complex and informal vocabulary but the Inter-quartile range for the tabloid will be greater but further down the graph. Therefore, as readability is considered I think a tabloid will be easier to read than a broadsheet, as the word length would be longer in a broadsheet, which uses more complex and informal language, which you would expect to have a longer word length.
From the data gathered, my hypothesis seems to be correct. In all three articles looked at, I found that the mean word length was greater in the broadsheet than the tabloid newspaper by around 0.3 letters. Here is a summary of the data gathered.
Article 1 - Front Page
Article 2 - Olympic Bid
Article 3 - Liverpool Manager
Average
Newspaper
Mean
I.Q Range
Median
Mean
I.Q. Range
Median
Mean
I.Q. Range
Median
Mean
I.Q Range
Median
The Times
4.7179
3.5
3.6
4.7692
2.6
4
4.1795
2.6
3.1
4.5556
2.9
3.5667
The Sun
4.2857
3.5
2.9
4.45
4.6
3.8
3.85
3.2
3.1
4.1952
3.76667
3.2667
Difference
0.4322
0
0.7
0.3192
-2
0.2
0.3295
-0.6
0
0.3603
-0.8667
0.3
As stated in my hypothesis, the inter quartile range of the word length in the tabloid is larger than in the broadsheet newspaper on average 0.867 larger. This is because the language used in a tabloid newspaper is far more informal and easier to read meaning the word length would be smaller. The mean and median were greater in the broadsheet than the tabloid articles. The mean had an average difference of 0.3603 and the median had an average difference of 0.3 between both newspapers.
The inter-quartile range in article 1 shows no difference between the newspapers however looking at the graphs it can be seen that the tabloid article has a smaller lower quartile value than the broadsheet. This fact is repeated in all the other articles too, i.e. there are smaller words in the tabloid newspaper than the broadsheet.
Newspaper
Article
Q1 Lower Quartile
Q2 Median
Q3 Upper Quartile
Inter Quartile Range
Tabloid
Front Page
.9
2.9
5.4
3.5
Broadsheet
Front Page
2.1
3.6
5.6
3.5
Tabloid
Olympic Bid
.8
3.8
5.4
3.6
Broadsheet
Olympic Bid
2.9
4
5.5
2.6
Tabloid
Liverpool Manager
.5
3.1
4.7
3.2
Broadsheet
Liverpool Manager
2
3.1
4.6
2.6
The distribution of data is shown in box plots (box and whisker diagrams). They show the position and size of the inter quartile range and the whiskers normally show the range of the data. The whiskers will show the range excluding what it sees at exceptional pieces of data in the case where there are outliers.
Firstly, the tabloid articles had no outliers but a larger range than the broadsheets (indicated by long whiskers). For the Liverpool manager third tabloid article, the range was 9 as apposed to the broadsheet article, which was only 8. Similarly, the Olympic bid tabloid article had a range of 12 but the broadsheet only had a range of 11. This data clearly shows that in the tabloid newspaper, the word lengths are more spread out. This suggests that the tabloid's mean word length has been raised by a spread of larger words; the whiskers on the right are much larger.
Tabloids have a fewer longer words which are spread out over a larger range. Unlike the broadsheets, that has a concentration of very long words that will be hard to read. The words in a tabloid are long are not as long as those in a broadsheet, however are easier to read. Although the mean of the Tabloid are very similar and not much different compared to the mean of the broadsheet.
The limitations to this investigation were that the actual articles in the front page of both the newspapers were different. This may have influenced the results obtained which may have lead to an error in my investigation, as perhaps the topic in one article required smaller words in order for it to be expressed than the other article topic giving bias results. Next time, it would be preferential to choose two newspapers with the same story on the front page. The tedious task of obtaining the word length may have caused some human errors resulting in another limitation. To improve this situation, perhaps a computer could be used to calculate the word length if the selected articles were scanned into it. The sample chosen was 40 this proved to be a reasonably sized sample allowing me to get a range of word sizes from the articles. A larger sample would lead to more anomalies in my investigation however would mean I would have a larger range and a more accurate measure. A smaller sized sample would have resulted me in fewer anomalies but would not have given me a range of word lengths throughout the whole article.
Percentage of Text to the Whole Article
In my hypothesis, I predicted that there would be greater percentage of writing compared to the picture in a broadsheet than a tabloid, because broadsheet papers write in more depth and therefore need more space for text rather than photos. I also thought the ratios would differ depending on the article topic and where it is in the paper.
My hypothesis proved correct and incorporating my results of both the tests carried out both proves to conclude the same thing that Broadsheets are easier to read than tabloids. Looking at the results gathered from the percentage of text to the rest of the article all three of the articles from each paper showed the same results, that tabloids have a lower text to rest of the article ratio and broadsheets have a higher text to rest of the article ratio. This shows that tabloids are easier to read than tabloids because broadsheet newspapers write in more depth and therefore need more space for text rather than photos.
To further investigate this data-handling project, a wider range of newspapers can be chosen to compare their average word length. It would establish whether the results of this investigation are a one off coincidence or whether all tabloids and broadsheets are typical to some extent. Another extension of this topic can be to compare whether the average word length changes depending on the day it is published. Weekend and weekday newspapers may vary within the tabloids and broadsheets. Other methods of readability can be used to compare the tabloids and broadsheet such as sentence length, size of paragraph, length of the article or size of writing. These may all extend this investigation.
Mathematics Coursework - Data Handling
Shaan Patel
